INDEX

Explanations

I think

This neuron fires on the very first word of a sentence—particularly common sentence-initial pronouns and discourse markers (e.g. “I,” “As,” “If,” “Don’t,” “The,” “So,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

in

-2.09

to

-1.81

or

-1.76

and

-1.72

an

-1.70

by

-1.70

 .....

-1.59

 Numerous

-1.57

all

-1.57

new

-1.56

POSITIVE LOGITS

 utiliser

1.99

 résister

1.95

 accompagner

1.95

镲

1.93

 emporter

1.93

牐

1.91

桖

1.91

didn

1.88

鼱

1.86

 envoyer

1.85

Activations Density 0.206%